Student Learning Heterogeneity in School Mathematics: A Comparison of Methods

• In Event: Poster Session 11
  In Poster Session: Assessment and Teacher Knowledge in Math

Abstract

The purpose of this study is to compare cut-score and latent class analysis methods when characterizing learning heterogeneity in school mathematics. Approximately 120 000 student responses to an Ontario knowledge and skills assessment instrument were gathered in grade 6 and linked to responses gathered from the same students three years later in grade 9. This jurisdiction uses an item responses modeling approach with cut-scores to determine appropriate student ability levels. I re-analyzed 36 grade 6 items (28 MC, 8 OR) and 32 grade 9 items (24 MC, 8OR) using a latent class analysis (within grades) and a latent transition analysis (between grades) approach. These results raise questions about the veracity of traditional cut-score based interpretations of learning differences.

Author

• Malcolm Cunningham, University of Toronto